“The Skynet funding bill is passed. The system goes online on August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29. In a panic, they try to pull the plug.”
The Terminator
Arnold Schwarzenegger
That sounded comfortably like science fiction when Terminator 2 came out in 1991. It reads a little differently today.
During September, leaders including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Elon Musk and Google DeepMind CEO Demis Hassabis were publicly warning about how quickly AI is advancing and how difficult the most capable systems could eventually become to control. Researchers have left some of the leading AI companies over these concerns, with one estimating the chance that AI ultimately causes human extinction at 10%. Alarmingly, models are sometimes behaving in ways their creators did not necessarily anticipate or intend.
The concern is what happens as these systems become increasingly capable of working on their own, solving problems, writing software and eventually helping to improve the next generation of AI. At some point, the speed at which the technology advances could begin to exceed our ability to understand or manage it.
That is an unsettling thought. What makes the moment even stranger is that, while many of the people building AI are warning us to be careful, companies around the world are spending extraordinary amounts of money to build more of it.
Of course, there is a pretty obvious reason why vast sums are being thrown at AI. AI adoption is running far ahead of the early adoption of previous technologies like personal computers and the internet. Something that barely existed as a consumer product a few years ago is now being used daily by the masses. The infrastructure required to support that growth is enormous. Current estimates put average annual AI infrastructure spending from 2025 through 2032 at roughly 3.6% of U.S. GDP. For some historical perspective, railroad investment during its great expansion averaged about 2.2% of GDP annually, while the highway buildout and the telecom and fiber boom were each closer to 1%.
This massive investment spending reaches far beyond technology companies. Data centers need semiconductors and servers, but they also need electricity, transmission capacity, land, construction, cooling equipment and financing. You end up with a huge amount of capital moving through the economy, and increasingly that is showing up in the economic data. Riding the AI wave, U.S. business activity has recently accelerated to its strongest level in five years. More than six years into this expansion, the economy has withstood a surge in inflation, sharply higher interest rates and one of the most aggressive Federal Reserve tightening cycles in decades. AI spending has become a big reason the economy still has this much momentum.
The strength in the economy has shown up just as clearly in corporate profits. Earnings have been a major driver of the stock market’s gains this year. Usually, Wall Street begins the year with optimistic earnings estimates and spends the next several quarters gradually cutting them. In 2026, the opposite has happened. Estimates have risen sharply as the year has progressed, with expectations for 2027 profits moving higher still. Roughly 90% of U.S. companies have beaten earnings expectations this year, the best rate in more than 20 years, and investors now expect long-term earnings growth for the S&P 500 at levels we have not seen in more than 25 years.
Today’s earnings picture helps explain why investors are willing to pay high valuations. If earnings are going to grow at 15%, 20% or more for a sustained period of time, a high P/E ratio becomes much easier to live with. Thus far, companies have given investors plenty of reasons to believe the growth is real.
We worry that a lot is being asked of earnings from here. Over most of the past 90 years, S&P 500 earnings have grown within a consistent range. Recent results have moved well above the top of that range. Corporate profits are also running at an all-time high share of the economy. Maybe some of that reflects a permanent change in the mix of the S&P 500 toward more profitable businesses. Eventually, however, unusually attractive economics tend to invite competition, leading to excess investment and higher costs. Growth is harder to achieve when starting from such a high base.
Companies tied directly to the AI infrastructure boom are producing much stronger earnings growth than the rest of the market, and those same companies are spending heavily to keep the buildout going. Based on operating cash flow before capital spending, valuations really do not appear that extreme. But once you deduct all the money being thrown at data centers, chips and related infrastructure, the picture changes. On a free cash flow basis, which measures the cash companies have left after those capital expenditures, the S&P 500 trades at close to a 20-year high.
Investors are assuming this level of AI spending will eventually produce very large profits. It is true that AI adoption is moving incredibly quickly, the technology is advancing at a remarkable rate, and the largest companies making these investments have the balance sheets to fund them. What we don’t know is whether the returns on investment will justify the scale of spending on a nascent, rapidly changing technology. With today’s outsized profit growth expected to continue apace, the margin for disappointment appears very thin.
The stock market’s blistering pace has rewarded investors handsomely. The S&P 500 has produced one of its strongest four-year stretches in a century. Underneath the index, though, the experience has been much more uneven. Over the past three years, only about 30% of stocks have managed to outperform the S&P 500. A concentrated group of very large companies delivered results strong enough to lift the cap-weighted S&P 500 Index far ahead of most individual stocks.
The stock market’s bifurcated returns can also be seen across sectors. Technology companies now trade at a substantial valuation premium vs. more defensive sectors such as health care, consumer staples and utilities. The only time that gap was wider was around the peak of the dot-com boom.
Narrow markets always make us a little uncomfortable. At the same time, we recognize an interesting characteristic of how investing in the stock market tends to work. Looking all the way back to 1925, roughly half of individual stocks actually lost money over their full lifetimes, while a relatively small number of enormous winners delivered most of the stock market’s total wealth creation. There have always been many more average businesses than truly exceptional ones.
The past few years’ narrow market environment leaves us pleased with how our individual stock portfolios have performed. We certainly do not expect every stock we own to be one of those generational winners, but this period has been a good reminder of how valuable it can be to own the right businesses when market returns are driven by only a handful of stocks.
The amount of money going into AI matters beyond earnings and stock prices. Somebody has to finance the buildout. The biggest spenders (“hyperscaler” tech companies) are starting to commit beyond their organic cash flows. Meanwhile, all this demand for capital is arriving at the same time the federal government is borrowing heavily and oil prices remain elevated, putting additional upward pressure on long-term interest rates. The 30-year Treasury now yields roughly 5.6%, its highest level in more than 20 years.
This year’s pronounced rise in interest rates has been brutal for investors who owned long-duration bonds. Treasuries with maturities of 10 years or longer are down materially this year even after accounting for their coupon income, adding to what has already been one of the worst five and ten year stretches for Treasury investors in history.
We were worried about this setup last year. With long-term yields still relatively low, we did not think investors were being paid enough to take as much interest-rate risk as the market was offering. We shortened maturities across portfolios and reduced exposure to securities that would be especially vulnerable if long rates moved higher, including our perpetual preferred stocks. We obviously did not know precisely where rates would go, but the risk/reward looked poor enough that we wanted less exposure to it. That decision has helped, as longer-duration bonds have continued to struggle, with negative total returns this year.
After years of pain in the bond market, the starting point today is much better. A Treasury yielding more than 5%, with no credit risk and interest exempt from state income tax, is a very different proposition than a Treasury yielding 2%. High-quality corporate bonds and other fixed-income securities offer even better yields that can make a meaningful contribution to portfolio returns without forcing investors far down the credit spectrum. Agency-backed mortgage securities are one example we like today. They offer additional yield over comparable Treasuries while retaining very high credit quality, with most of that extra compensation coming from mortgage prepayment and extension risk rather than credit risk.
One of the primary asset allocation decisions is stocks vs. bonds. For years, low interest rates made the decision relatively easy. Safe bonds paid next to nothing, so investors wanting a reasonable return had little choice but to take more risk. At 5% or better, fixed income is finally legitimate competition for stocks.
We remain very comfortable owning stocks. The economy is strong, earnings are excellent and there is a reasonable chance AI turns out to be every bit as profitable as the spending around it suggests. We are also mindful, however, that we are coming off one of the strongest runs in stock market history, earnings expectations are very high, and the companies carrying the market are being asked to justify an enormous amount of investment.
Fortunately, we no longer need the S&P 500 Index to do all of the work. Dispersion underneath the index is creating opportunities away from the most obvious recent winners, and higher bond yields give us another attractive source of return. That makes a strong case for staying invested while being thoughtful about where we are taking risk.
Thank you for your continued trust and support.
Sincerely,
John G. Prichard Kurt Beimfohr Miles E. Yourman
Past performance is not indicative of future results. The above information is based on internal research derived from various sources and does not purport to be a statement of all material facts relating to the information and markets mentioned. It should not be construed that the information in this commentary is a recommendation to purchase or sell any securities.
Fall 2026 Quarterly Commentary
October 9th, 2026
The Terminator
Arnold Schwarzenegger
That sounded comfortably like science fiction when Terminator 2 came out in 1991. It reads a little differently today.
During September, leaders including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Elon Musk and Google DeepMind CEO Demis Hassabis were publicly warning about how quickly AI is advancing and how difficult the most capable systems could eventually become to control. Researchers have left some of the leading AI companies over these concerns, with one estimating the chance that AI ultimately causes human extinction at 10%. Alarmingly, models are sometimes behaving in ways their creators did not necessarily anticipate or intend.
The concern is what happens as these systems become increasingly capable of working on their own, solving problems, writing software and eventually helping to improve the next generation of AI. At some point, the speed at which the technology advances could begin to exceed our ability to understand or manage it.
That is an unsettling thought. What makes the moment even stranger is that, while many of the people building AI are warning us to be careful, companies around the world are spending extraordinary amounts of money to build more of it.
Of course, there is a pretty obvious reason why vast sums are being thrown at AI. AI adoption is running far ahead of the early adoption of previous technologies like personal computers and the internet. Something that barely existed as a consumer product a few years ago is now being used daily by the masses. The infrastructure required to support that growth is enormous. Current estimates put average annual AI infrastructure spending from 2025 through 2032 at roughly 3.6% of U.S. GDP. For some historical perspective, railroad investment during its great expansion averaged about 2.2% of GDP annually, while the highway buildout and the telecom and fiber boom were each closer to 1%.
This massive investment spending reaches far beyond technology companies. Data centers need semiconductors and servers, but they also need electricity, transmission capacity, land, construction, cooling equipment and financing. You end up with a huge amount of capital moving through the economy, and increasingly that is showing up in the economic data. Riding the AI wave, U.S. business activity has recently accelerated to its strongest level in five years. More than six years into this expansion, the economy has withstood a surge in inflation, sharply higher interest rates and one of the most aggressive Federal Reserve tightening cycles in decades. AI spending has become a big reason the economy still has this much momentum.
The strength in the economy has shown up just as clearly in corporate profits. Earnings have been a major driver of the stock market’s gains this year. Usually, Wall Street begins the year with optimistic earnings estimates and spends the next several quarters gradually cutting them. In 2026, the opposite has happened. Estimates have risen sharply as the year has progressed, with expectations for 2027 profits moving higher still. Roughly 90% of U.S. companies have beaten earnings expectations this year, the best rate in more than 20 years, and investors now expect long-term earnings growth for the S&P 500 at levels we have not seen in more than 25 years.
Today’s earnings picture helps explain why investors are willing to pay high valuations. If earnings are going to grow at 15%, 20% or more for a sustained period of time, a high P/E ratio becomes much easier to live with. Thus far, companies have given investors plenty of reasons to believe the growth is real.
We worry that a lot is being asked of earnings from here. Over most of the past 90 years, S&P 500 earnings have grown within a consistent range. Recent results have moved well above the top of that range. Corporate profits are also running at an all-time high share of the economy. Maybe some of that reflects a permanent change in the mix of the S&P 500 toward more profitable businesses. Eventually, however, unusually attractive economics tend to invite competition, leading to excess investment and higher costs. Growth is harder to achieve when starting from such a high base.
Companies tied directly to the AI infrastructure boom are producing much stronger earnings growth than the rest of the market, and those same companies are spending heavily to keep the buildout going. Based on operating cash flow before capital spending, valuations really do not appear that extreme. But once you deduct all the money being thrown at data centers, chips and related infrastructure, the picture changes. On a free cash flow basis, which measures the cash companies have left after those capital expenditures, the S&P 500 trades at close to a 20-year high.
Investors are assuming this level of AI spending will eventually produce very large profits. It is true that AI adoption is moving incredibly quickly, the technology is advancing at a remarkable rate, and the largest companies making these investments have the balance sheets to fund them. What we don’t know is whether the returns on investment will justify the scale of spending on a nascent, rapidly changing technology. With today’s outsized profit growth expected to continue apace, the margin for disappointment appears very thin.
The stock market’s blistering pace has rewarded investors handsomely. The S&P 500 has produced one of its strongest four-year stretches in a century. Underneath the index, though, the experience has been much more uneven. Over the past three years, only about 30% of stocks have managed to outperform the S&P 500. A concentrated group of very large companies delivered results strong enough to lift the cap-weighted S&P 500 Index far ahead of most individual stocks.
The stock market’s bifurcated returns can also be seen across sectors. Technology companies now trade at a substantial valuation premium vs. more defensive sectors such as health care, consumer staples and utilities. The only time that gap was wider was around the peak of the dot-com boom.
Narrow markets always make us a little uncomfortable. At the same time, we recognize an interesting characteristic of how investing in the stock market tends to work. Looking all the way back to 1925, roughly half of individual stocks actually lost money over their full lifetimes, while a relatively small number of enormous winners delivered most of the stock market’s total wealth creation. There have always been many more average businesses than truly exceptional ones.
The past few years’ narrow market environment leaves us pleased with how our individual stock portfolios have performed. We certainly do not expect every stock we own to be one of those generational winners, but this period has been a good reminder of how valuable it can be to own the right businesses when market returns are driven by only a handful of stocks.
The amount of money going into AI matters beyond earnings and stock prices. Somebody has to finance the buildout. The biggest spenders (“hyperscaler” tech companies) are starting to commit beyond their organic cash flows. Meanwhile, all this demand for capital is arriving at the same time the federal government is borrowing heavily and oil prices remain elevated, putting additional upward pressure on long-term interest rates. The 30-year Treasury now yields roughly 5.6%, its highest level in more than 20 years.
This year’s pronounced rise in interest rates has been brutal for investors who owned long-duration bonds. Treasuries with maturities of 10 years or longer are down materially this year even after accounting for their coupon income, adding to what has already been one of the worst five and ten year stretches for Treasury investors in history.
We were worried about this setup last year. With long-term yields still relatively low, we did not think investors were being paid enough to take as much interest-rate risk as the market was offering. We shortened maturities across portfolios and reduced exposure to securities that would be especially vulnerable if long rates moved higher, including our perpetual preferred stocks. We obviously did not know precisely where rates would go, but the risk/reward looked poor enough that we wanted less exposure to it. That decision has helped, as longer-duration bonds have continued to struggle, with negative total returns this year.
After years of pain in the bond market, the starting point today is much better. A Treasury yielding more than 5%, with no credit risk and interest exempt from state income tax, is a very different proposition than a Treasury yielding 2%. High-quality corporate bonds and other fixed-income securities offer even better yields that can make a meaningful contribution to portfolio returns without forcing investors far down the credit spectrum. Agency-backed mortgage securities are one example we like today. They offer additional yield over comparable Treasuries while retaining very high credit quality, with most of that extra compensation coming from mortgage prepayment and extension risk rather than credit risk.
One of the primary asset allocation decisions is stocks vs. bonds. For years, low interest rates made the decision relatively easy. Safe bonds paid next to nothing, so investors wanting a reasonable return had little choice but to take more risk. At 5% or better, fixed income is finally legitimate competition for stocks.
We remain very comfortable owning stocks. The economy is strong, earnings are excellent and there is a reasonable chance AI turns out to be every bit as profitable as the spending around it suggests. We are also mindful, however, that we are coming off one of the strongest runs in stock market history, earnings expectations are very high, and the companies carrying the market are being asked to justify an enormous amount of investment.
Fortunately, we no longer need the S&P 500 Index to do all of the work. Dispersion underneath the index is creating opportunities away from the most obvious recent winners, and higher bond yields give us another attractive source of return. That makes a strong case for staying invested while being thoughtful about where we are taking risk.
Thank you for your continued trust and support.
Sincerely,
John G. Prichard Kurt Beimfohr Miles E. Yourman
Past performance is not indicative of future results. The above information is based on internal research derived from various sources and does not purport to be a statement of all material facts relating to the information and markets mentioned. It should not be construed that the information in this commentary is a recommendation to purchase or sell any securities.